Where Will the Jobs Be?
It’s usually two steps forward, one step back, but the trend has been forward.
Posted Sep 07, 2017
The University of California, Berkeley has invited me to give a lecture open to the public. I’m calling it "The Future of Work." Here is its key content.
Indeed the world of work is a-changin': changes because of globalization, which party is in office, and probably most potent, automation. That’s not just “The Robots are Coming, The Robots are Coming” but because of ever more capable computers, increasingly self-teaching, so-called machine learning, deep learning, or artificial intelligence.
Of course, he who lives by the crystal ball eats broken glass. To reduce that risk, my predictions here are based on centrist assumptions: where there’s reasonable consensus or where logic makes a prediction low-risk, for example, an extension of an existing trend.
What jobs are going away?
One such trend is our preference for automation that will save us time: Most of us prefer an ATM to a teller, drive-through tolls to a toll-taker, buying on Amazon to braving traffic and parking hassles. And many of us wouldn't lament replacing the typical retail clerk with an all-knowing robot. By the way, that’s not science fiction, Lowe’s Home Improvement Centers are trialing Lowebot. You can, for example, put a bolt into Lowebot's opening and it will tell you, in any of 12 languages, what the matching nut is and walk you to it.
Another such trend is robotization of manufacture. For the example, the Tesla factory floor is filled mainly with robotics, not people. Many people feel it’s “social justice" for employers to pay Social Security, Workers Comp, Paid Family Leave, ObamaCare etc. Alas, that makes it ever more costly to hire an American, so employers are ever more tempted to automate if only to avoid going out of business, losing to low-labor-cost countries. And it's hard to argue that for routine tasks, people are more reliable than robots. For example, robots never come in hung-over, so they produce no "Monday morning cars." Nor are robots prone to human frailties. An assembly-line worker at a Ford plant told me that the workers pull pranks such as dropping a bolt into a car’s rear end to befuddle the quality-control person trying to find the cause. A robot wouldn’t do that.
Employers and the public generally also welcome machines doing work that’s mind-numbingly repetitive such as backbreaking farming. Japan already has an almost completely people-free farm. Self-driving trucks and trains, still probably a decade away, will eliminate jobs that are dull yet require perfect attention, for example, driving a truck or freight train cross-country. Exhausting warehouse jobs and dangerous ones such as soldier will be at least partly replaced with robots such as this one.
What percent of jobs are vulnerable to automation? Estimates for the next 20 years range from 9% to 47%. Of course, routine jobs are most vulnerable but so are some not-routine jobs: IBM’s Watson is making as accurate cancer diagnosis as doctors’. Law firms are using computers to replace lawyers in doing document review and discovery—the gathering of information on a case. Many of us use TurboTax to do our tax returns when we used to hire an accountant. Now, we may only ask the accountant for strategic advice or to review our returns—far fewer accountants needed.
Where will the jobs be?
Of course, that leads us to that more emotion-loaded question: What good jobs will be left for humans? Which careers will be automation-resistant, offshore resistant, and obsolescence resistant? Think of those buggy whip makers after cars came along.
Four categories of jobs should remain robust:
- Requiring ability to interpret data, for example, program evaluators, market strategists, and government program analysts.
- Requiring understanding of human subtleties: for example, salespeople of complex items, fundraisers, people managers, and counselors.
- Hands-on: e.g., intermediate health care providers such as physician assistants and physical therapy assistants, and tech installers and repairers, for example, mobile autonomous vehicle servicers.
- Simple local microbusiness owners, for example, tutor, flower cart owner, and medical advocate who, for example, helps newly diagnosed cancer patients negotiate the health care system.
Let’s move from the micro to the macro.
A mainstream pessimistic scenario
In the past, technology ended up creating more jobs than it eliminated. But this time may be different because so much of what we produce will be created by ever smarter machines, so-called self-teaching/deep learning computers. Assuming that median estimate of 1/3 job-loss in 20 years, such a rise in unemployment and underemployment would cause an increase in unhappiness, drug abuse, and civil unrest, and increased probability of wars, powerful ones because of advances in technology.
A mainstream optimistic scenario
Such a scenario predicts that, as has always been the case, technology will, net, create jobs or at least not cause significant job loss. And advances in education—for example, dream-team taught, highly immersive learning modules available to rich and poor worldwide—will yield a more educated and career-ready populus. Also, society will develop innovative job creation strategies. Here are two I’ve come up with:
An Assistance Army: The public would be encouraged to hire personal assistants: new-parent helper, school tutor, tech tutor, personal assistant, elder companion. That would create humanly if not financially rewarding jobs.
An Entrepreneurship Army: From kindergarten lemonade class projects to high school entrepreneurship courses to midlife adult job retraining, society would create many more entrepreneurs. Entrepreneurship is the only job creation engine that, unlike government employment, doesn’t require eating our seed corn
Whatever decline in income that would accrue from technology-caused loss of jobs would be replaced by increased valuing of creative outlets and relationships, and decreased materialism, which the schools, colleges and media will likely encourage. In turn, that would improve our environment—less stuff would be made. The environment would additionally be improved because technology will create cheaper better solar, wind and safer nuclear. Indeed, no less than Bill Gates is investing in small much safer nuclear plants.
Computers will facilitate ever better medical research, curing diseases faster, maybe even preventive gene therapy to guarantee we don’t get cancer. If society deems it wise, we may also be able to gene-edit our sperm and eggs so our kids have the genetic predisposition to altruism and good reasoning ability. While humans will always be needed as caretakers, computers will help. For example, see this remarkable combination health coach and high-tech medicine dispenser and reminder. Computer-driven decisions may be more ethical, for example, in sales, hiring and claims-paying.
What do I predict? I prefer to bet on the optimistic side. For millennia, there have been naysayers but while it’s usually two steps forward, one step back, the trend has been inexorably forward.
The lecture has now occurred. Here's the audio.